Correlation set analysis: detecting active regulators in disease populations using prior causal knowledge.

Area of Science:

• Systems biology
• Computational biology
• Genomics

Background:

• Identifying active causal regulators is key for understanding disease mechanisms and discovering drug targets.
• Existing methods often require large sample sizes or diverse data types.
• Integrating prior biological knowledge enhances the accuracy of regulator identification.

Purpose of the Study:

• To present a novel data-driven method, Correlation Set Analysis (CSA), for detecting active regulators in disease populations.
• To demonstrate CSA's ability to integrate co-expression analysis with literature-derived causal relationships.
• To provide a tool for uncovering both known and novel active regulators.

Main Methods:

• Correlation Set Analysis (CSA) integrates gene co-expression analysis with literature-derived causal networks.
• Focuses on the coherence of a regulator's target genes (regulatees) rather than direct co-expression.
• Utilizes simulated and real biological datasets for validation.

Main Results:

• CSA effectively recovers weak regulatory relationships with a low false discovery rate on simulated data.
• Successfully identified known and novel active regulators across three distinct real disease datasets.
• The method reveals both single and higher-order regulatory interactions.

Conclusions:

• CSA offers an intuitive, data-driven approach to identify disease-relevant regulators.
• The method facilitates the selection of targeted perturbation experiments for further investigation.
• Combining co-expression analysis with prior biological networks successfully identifies causal regulators and aids in formulating disease progression hypotheses.